Extensive surveys of the fatty acid composition of seed oils from different species of higher plants have resulted in the identification of at least 33 structurally distinct monohydroxylated plant fatty acids, and 12 different polyhydroxylated fatty acids that are accumulated by one or more plant species (reviewed by van de Loo et al. 1993). Ricinoleic acid, the principal constituent of the seed oil from the castor plant Ricinus communis (L.), is of commercial importance. We have previously described the cloning of a gene from this species that encodes a fatty acid hydroxylase, and the use of this gene to produce ricinoleic acid in transgenic plants of other species (see U.S. patent application Ser. No. 08/320,982, filed Oct. 11, 1994). The scientific evidence supporting the claims in that patent application were subsequently published (van de Loo et al., 1995).
The use of the castor hydroxylase gene to also produce other hydroxylated fatty acids such as lesquerolic acid, densipolic acid, hydroxypalmitoleic, hydroxyerucic and auricolic acid in transgenic plants is the subject of this invention. In addition, the identification of a gene encoding a homologous hydroxylase from Lesquerella fendleri, and the use of this gene to produce these hydroxylated fatty acids in transgenic plants is the subject of this invention.
Castor is a minor oilseed crop. Approximately 50% of the seed weight is oil (triacylglycerol) in which 85-90% of total fatty acids are the hydroxylated fatty acid, ricinoleic acid. Oil pressed or extracted from castor seeds has many industrial uses based upon the properties endowed by the hydroxylated fatty acid. The most important uses are production of paints and varnishes, nylon-type synthetic polymers, resins, lubricants, and cosmetics (Atsmon 1989).
In addition to oil, the castor seed contains the extremely toxic protein ricin, allergenic proteins, and the alkaloid ricinine. These constituents preclude the use of the untreated seed meal (following oil extraction) as a livestock feed, normally an important economic aspect of oilseed utilization. Furthermore, with the variable nature of castor plants and a lack of investment in breeding, castor has few favorable agronomic characteristics.
For a combination of these reasons, castor is no longer grown in the United States and the development of an alternative domestic source of hydroxylated fatty acids would be attractive. The production of ricinoleic acid, the important constituent of castor oil, in an established oilseed crop through genetic engineering would be a particularly effective means of creating a domestic source.
Because there is no practical source of lesquerolic, densipolic and auricolic acids from plants that are adapted to modern agricultural practices, there is currently no large-scale use of these fatty acids by industry. However, the fatty acids would have uses similar to those of ricinoleic acid if they could be produced in large quantities at comparable cost to other plant-derived fatty acids (Smith 1985).
Plant species, such as certain species in the genus Lesquerella, that accumulate a high proportion of these fatty acids, have not been domesticated and are not currently considered a practical source of fatty acids (Hirsinger, 1989). This invention represents a useful step toward the eventual production of these and other hydroxylated fatty acids in transgenic plants of agricultural importance.
The taxonomic relationships between plants having similar or identical kinds of unusual fatty acids have been examined (van de Loo et al., 1993). In some cases, particular fatty acids occur mostly or solely in related taxa. In other cases there does not appear to be a direct link between taxonomic relationships and the occurrence of unusual fatty acids. In this respect, ricinoleic acid has now been identified in 12 genera from 10 families (reviewed in van de Loo et al., 1993). Thus, it appears that the ability to synthesize hydroxylated fatty acids has evolved several times independently during the radiation of the angiosperms. This suggested to us that the enzymes which introduce hydroxyl groups into fatty acids arose by minor modifications of a related enzyme.
Indeed, as shown herein, the sequence similarity between Δ12 fatty acid desaturases and the kappa hydroxylase from castor is so high that it is not possible to unambiguously determine whether a particular enzyme is a desaturase or a hydroxylase on the basis of evidence in the scientific literature. Similarly, a patent application (PCT/US93/09987) that purports to teach the isolation and use of Δ12 fatty acid desaturases does not teach how to distinguish a hydroxylase from a desaturase. In view of the importance of being able to distinguish between these activities for the purpose of genetic engineering of plant oils, the utility of that application is limited to the several instances where direct experimental evidence (e.g., altered fatty acid composition in transgenic plants) was presented to support the assignment of function. A method for distinguishing between fatty acid desaturases and fatty acid hydroxylases on the basis of amino acid sequence of the enzyme is also a subject of this invention.
A feature of hydroxylated or other unusual fatty acids is that they are generally confined to seed triacylglycerols, being largely excluded from the polar lipids by unknown mechanisms (Battey and Ohlrogge 1989; Prasad et al., 1987). This is particularly intriguing since diacylglycerol is a precursor of both triacylglycerol and polar lipid. With castor microsomes, there is some evidence that the pool of ricinoleoyl-containing polar lipid is minimized by a preference of diacylglycerol acyltransferase for ricinoleate-containing diacylglycerols (Bafor et al. 1991). Analyses of vegetative tissues have generated few reports of unusual fatty acids, other than those occurring in the cuticle. The cuticle contains various hydroxylated fatty acids which are interesterified to produce a high molecular weight polyester which serves a structural role. A small number of other exceptions exist in which unusual fatty acids are found in tissues other than the seed.
The biosynthesis of ricinoleic acid from oleic acid in the developing endosperm of castor (Ricinus communis) has been studied by a variety of methods. Morris (1967) established in double-labeling studies that hydroxylation occurs directly by hydroxyl substitution rather than via an unsaturated-, keto- or epoxy-intermediate. Hydroxylation using oleoyl-CoA as precursor can be demonstrated in crude preparations or microsomes, but activity in microsomes is unstable and variable, and isolation of the microsomes involved a considerable, or sometimes complete loss of activity (Galliard and Stumpf, 1966; Moreau and Stumpf, 1981). Oleic acid can replace oleoyl-CoA as a precursor, but only in the presence of CoA, Mg2+ and ATP (Galliard and Stumpf, 1966) indicating that activation to the acyl-CoA is necessary. However, no radioactivity could be detected in ricinoleoyl-CoA (Moreau and Stumpf, 1981). These and more recent observations (Bafor et al., 1991) have been interpreted as evidence that the substrate for the castor oleate hydroxylase is oleic acid esterified to phosphatidyicholine or another phospholipid.
The hydroxylase is sensitive to cyanide and azide, and dialysis against metal chelators reduces activity, which could be restored by addition of FeSO4, suggesting iron involvement in enzyme activity (Galliard and Stumpf, 1966). Ricinoleic acid synthesis requires molecular oxygen (Galliard and Stumpf, 1966; Moreau and Stumpf 1981) and requires NAD(P)H to reduce cytochrome b5 which is thought to be the intermediate electron donor for the hydroxylase reaction (Smith et al., 1992). Carbon monoxide does not inhibit hydroxylation, indicating that a cytochrome P450 is not involved (Galliard and Stumpf, 1966; Moreau and Stumpf 1981). Data from a study of the substrate specificity of the hydroxylase show that all substrate parameters (i.e., chain length and double bond position with respect to both ends) are important; deviations in these parameters caused reduced activity relative to oleic acid (Howling et al., 1972). The position at which the hydroxyl was introduced, however, was determined by the position of the double bond, always being three carbons distal. Thus, the castor acyl hydroxylase enzyme can produce a family of different hydroxylated fatty acids depending on the availability of substrates. Thus, as a matter of convenience, we refer to the enzyme throughout as a kappa hydroxylase (rather than an oleate hydroxylase) to indicate the broad substrate specificity.
The castor kappa hydroxylase has many superficial similarities to the microsomal fatty acyl desaturases (Browse and Somerville, 1991). In particular, plants have a microsomal oleate desaturase active at the Δ12 position. The substrate of this enzyme (Schmidt et al., 1993) and of the hydroxylase (Bafor et al., 1991) appears to be a fatty acid esterified to the sn-2 position of phosphatidylcholine. When oleate is the substrate, the modification occurs at the same position (Δ12) in the carbon chain, and requires the same cofactors, namely electrons from NADH via cytochrome b5 and molecular oxygen. Neither enzyme is inhibited by carbon monoxide (Moreau and Stumpf, 1981), the characteristic inhibitor of cytochrome P450 enzymes.
There do not appear to have been any published biochemical studies of the properties of the hydroxylase enzyme(s) in Lesquerella. 
Conceptual Basis of the Invention
In U.S. patent application Ser. No. 08/320,982, we described the use of a cDNA clone from castor for the production of ricinoleic acid in transgenic plants. As noted above, biochemical studies by others had suggested that the castor hydroxylase may not have strict specificity for oleic acid but would also catalyze hydroxylation of other fatty acids such as icosenoic acid (20:1cisΔ11) (Howling et al., 1972). Based on these studies, our previous application Ser. No. 08/320,982 noted in Example 2 that the expression of the castor hydroxylase in transgenic plants of species such as Brassica napus and Arabidopsis thaliana that accumulate fatty acids such as icosenoic acid (20:1cisΔ11) and erucic acid (13-docosenoic acid; 22:1cisΔ13) would be expected to accumulate some of the hydroxylated derivatives of these fatty acids due to the activity of the hydroxylase on these fatty acids. We have now obtained additional direct evidence for such a claim based on the production of ricinoleic, lesquerolic, densipolic and auricolic fatty acids in transgenic Arabidopsis plants and have included such evidence herein as Example 1.
In Example 3 of the previous application, we taught the various methods by which the castor hydroxylase clone and sequences derived thereof could be used to identify other hydroxylase clones from plant species such as Lesquerella fendleri that are known to accumulate hydroxylated fatty acids in seed oils. In this continuation we have provided an example of the use of that aspect of the invention for the isolation of a novel hydroxylase gene from Lesquerella fendleri. 
In view of the high degree of sequence similarity between Δ12 fatty acid desaturases and the castor hydroxylase (van de Loo et al., 1995), the validity of claims (e.g., PCT WO 94/11516) for the use of desaturase or hydroxylase genes or sequences derived therefrom for the identification of genes of identical function from other species must be viewed with skepticism. In this application, we teach a method by which hydroxylase genes can be distinguished from desaturases and describe methods by which Δ12 desaturases can be converted to hydroxylases by the modification of the gene encoding the desaturases. A mechanistic basis for the similar reaction mechanisms of desaturases and hydroxylases was presented in the earlier patent application Ser. No. 08/320,982. Briefly, the available evidence suggests that fatty acid desaturases have a similar reaction mechanism to the bacterial enzyme methane monooxygenase which catalyses a reaction involving oxygen-atom transfer (CH4→CH3OH) (van de Loo et al., 1993). The cofactor in the hydroxylase component of methane monooxygenase is termed a μ-oxo bridged diiron cluster (FeOFe). The two iron atoms of the FeOFe cluster are liganded by protein-derived nitrogen or oxygen atoms, and are tightly redox-coupled by the covalently-bridging oxygen atom. The FeOFe cluster accepts two electrons, reducing it to the diferrous state, before oxygen binding. Upon oxygen binding, it is likely that heterolytic cleavage also occurs, leading to a high valent oxoiron reactive species that is stabilized by resonance rearrangements possible within the tightly coupled FeOFe cluster. The stabilized high-valent oxoiron state of methane monooxygenase is capable of proton extraction from methane, followed by oxygen transfer, giving methanol. The FeOFe cofactor has been shown to be directly relevant to plant fatty acid modifications by the demonstration that castor stearoyl-ACP desaturase contains this type of cofactor (Fox et al., 1993).
On the basis of the foregoing considerations, we hypothesized that the castor oleate hydroxylase is a structurally modified fatty acyl desaturase, based upon three arguments. The first argument involves the taxonomic distribution of plants containing ricinoleic acid. Ricinoleic acid has been found in 12 genera of 10 families of higher plants (reviewed in van de Loo et al., 1993). Thus, plants in which ricinoleic acid occurs are found throughout the plant kingdom, yet close relatives of these plants do not contain the unusual fatty acid. This pattern suggests that the ability to synthesize ricinoleic acid has arisen (and been lost) several times independently, and is therefore a quite recent divergence. In other words, the ability to synthesize ricinoleic acid has evolved rapidly, suggesting that a relatively minor genetic change in the structure of the ancestral enzyme was necessary to accomplish it.
The second argument is that many biochemical properties of castor kappa hydroxylase are similar to those of the microsomal desaturases, as discussed above (e.g., both preferentially act on fatty acids esterified to the sn-2 position of phosphatidylcholine, both use cytochrome b5 as an intermediate electron donor, both are inhibited by cyanide, both require molecular oxygen as a substrate, both are thought to be located in the endoplasmic reticulum).
The third argument stems from the discussion of oxygenase cofactors above, in which it is suggested that the plant membrane bound fatty acid desaturases may have a μ-oxo bridged diiron cluster-type cofactor, and that such cofactors are capable of catalyzing both fatty acid desaturations and hydroxylations, depending upon the electronic and structural properties of the protein active site.
Taking these three arguments together, it was hypothesized that kappa hydroxylase of castor endosperm is homologous to the microsomal oleate Δ12 desaturase found in all plants. The evidence supporting this hypothesis was disclosed in the previous patent application Ser. No. 08/320,982. A number of genes encoding microsomal Δ12 desaturases from various species have recently been cloned (Okuley et al., 1994) and substantial information about the structure of these enzymes is now known (Shanklin et al. 1994). Hence, in the following invention we teach how to use structural information about fatty acyl desaturases to isolate kappa hydroxylase genes of this invention. This example teaches the method by which any carbon-monoxide insensitive plant fatty acyl hydroxylase gene can be identified by one skilled in the art.
An unpredicted outcome of our studies on the castor hydroxylase gene in transgenic Arabidopsis plants was the discovery that expression of the hydroxylase leads to increased accumulation of oleic acid in seed lipids. Because of the low nucleotide sequence homology between the castor hydroxylase and the Δ12-desaturase (about 67%), we consider it unlikely that this effect is due to silencing (also called sense-suppression or cosuppression) of the expression of the desaturase gene by the hydroxylase gene. Whatever the basis for the effect, this invention teaches the use of hydroxylase genes to alter the level of fatty acid unsaturation in transgenic plants. On the basis of a hypothesis about the mechanisms of the effect, this invention also teaches the use of genetically modified hydroxylase and desaturase genes to achieve directed modification of fatty acid unsaturation levels.